• Front Page
   
  
• July ‘01 Issue
   
  
• April ‘01 Issue
   
  
• January ‘01 Issue
   
  
• October ‘00 Issue
   
  
• July ‘00 Issue
   
  
• April ‘00 Issue
   
  
• January ‘00 Issue
   
  
• October ‘99 Issue
   
  
• July ‘99 Issue
   
  
• April ‘99 Issue
   
  
• January ‘99 Issue
   
  
• October ‘98 Issue
   
  
• July ‘98 Issue
   
  
• April ‘98 Issue
   
  
• January ‘98 Issue
   
  
• October ‘97 Issue
   
  
• July ‘97 Issue
   
  

Choosing the Best Bid:

An Application of Two Managerial Decision-Making Theories

A Term Paper
Presented to the Department of Management
University of Nebraska
July 13, 2001

Freelance work provides a chance to those individuals who want to integrate their personal lives with their work without having to deal with organizational structures. It allows flexible work schedules. It gives them control over their earnings and their time. It allows them to choose who they are going to work with and what they will be doing.

Although freelance work has its own advantages, it also has some disadvantages. While working as a freelancer, the individual has to play the roles of both subordinate and manager. Managerial decisions have to be made by the same person who does the work. This is one of the most formidable challenges freelancers face in their work life because many of them have to rely on their hunches while making decisions if they have no managerial background or experience. On the other hand the applicability of many decision-making theories to freelance work has not yet been tested.

In this paper I apply the Game Theory and the Decision Theory principles to a freelance translation business and an online bidding situation.

Game Theory and Decision Theory could be considered two separate parts, or branches of the same decision-making theory. Both theories use mathematical and statistical methods in order to generate the necessary information for the decisions to be made. Game theory can be defined as a mathematical analysis, which is used in decision-making in conflict situations, whereas Decision Theory is generally more concerned about alternatives, natural states, and payoffs. Game Theory assumes that a conflict situation involves two or more decision-makers with different objectives, but "act on the same system or share the same resources" (Web Dictionary of Principa-Cybernatica). Decision theory, on the other hand, tries to provide information about which decisions should be made under certainty, risk, and uncertainty conditions.

In an article that was published online by Paul Walker in 1995, the history of Game theory is traced back to ancient times. According to Walker, the first examples of Game Theory could be found in 0-500 AD in the Babylonian Talmud. The Talmud is a book that contains the laws that constituted the basic of Jewish religious, criminal, and civil law. The regulations the Talmud brings to the distribution of heritage among the wives after the death of the husband were recognized by scholars in 1985 as the oldest sample of the modern theory of corporative games. According to the same article, the minimax mixed strategy solution was first introduced by James Waldegrave in 1713. Schniederjans (2001) states that many scholars credit J. Von Neumann and O. Morgenstren's book, "Theory of Games and Economic Behavior," published in 1944.

Game theory problems can have two or more players. According to Schniederjans (2001) Game Theory (GT) has two objectives: a) maximizing the payoffs of player one, b) minimizing the loss of player two. Thus, it could be argued that GT problems force the players to choose optimal compromised solutions. This must be one of the reasons why GT has been applied widely to many different areas of economic and business life. For example, in an article published by Roth (1999), GT is applied to market design, Dubois, Fargier, Pride (2000) discussed GT in decision making processes under ordinal preferences and uncertainty. Roth and Oscines (2001) discussed the changing bidding behaviors of online customers in eBay and Amazon.com. In an article published by Sudhir (2001) the applicability of GT theory in determining the competitive prices in the automobile industry is discussed. In another study Roth (2001) questioned the role of rationality in GT.

In this study both Game and Decision Theory principles are applied to an online freelance situation. Although there are studies published about the changing bidding habits of online customers (i.e., Roth and Ockeness, 2001) I was unable to find any article directly related to on-line freelance business and the application of these two theories to the decision-making processes of a freelance-based business.

The purpose of this paper is to discuss the applicability of the two theories mentioned above to a freelance translation business. The study is designed as a case study. The study discusses the applicability of the managerial decision-making theories to an online bidding situation and analyzes the ways the translation agencies make their decisions while choosing to whom the project should be granted. AB/QM software is used in order to analyze and interpret the data. AB/QM is a software package that is widely used in managerial decision-making processes.

The data for the case study is gathered from proz.com web site. Proz.com runs one of the on-line bidding bulletin boards for numerous translation agencies and freelance translators all over the world. A job announcement was posted on this site on June 19. This announcement and 27 bids it received in one day constituted the main data. Bids for this job were public. I assumed the role of the translation agency and tried to choose the best bid among 27 different bids by using decision theory principles.

The translation business forces translation agencies to depend on freelance translation providers. This is simply because they deal with many different language combinations every day and it is impossible (unfeasible) to have a professional translator from each and every language combination on staff. In time they create their own databases of translators and they establish long-term relationships with their best service providers in a number of language combinations. Although it is a fast-paced and very competitive market, once the relationship is established between the translator and the agency, those freelance translators who have managed to earn a good reputation with the agency are the first ones that are contacted when the need arises in their language combinations. These jobs are occasionally advertised on online bulletin boards. If a job is advertised online that means: a) The agency has no one in the required language combination with whom they have a long-term relationship; b) they are not happy with the one they used before; c) the translator they used to work with is not available; d) the agency is trying to make the largest profit possible by choosing the best, cheapest, and/or the fastest translator. Different combinations of these possibilities make agencies open the projects to bids. As a result, many freelance projects are posted online. After the freelance translators submit the bids, the translation agency makes its decision. As my three-year experience in the area showed me, the jobs that are advertised online receive countless bids from all over the world. The challenge a freelance translator faces becomes crucial: How can a translator make his/her bid stand out among those countless other offers? This is crucial because in the freelance translation business one of the most effective and cheapest ways of expanding one's client base is attracting new clients. Sudhir (2001) argues that one must assume an aggressive competitive stance in order to attract first-time buyers and thus achieve one's short-term goals. Competitive market rules also apply to this bidding game among translators, but, in contrast with Sudhir's (2001) statement, they affect both long-term and short-term situations of the freelance translator, because offering competitive prices not only helps translators accomplish their short-term goals by winning the bid, but also their long-term goals by expanding their client base. As a result, competitive pricing becomes the crucial element for freelance translators. Sudhir (2001) argues that Game and Decision Theories can be used in determining competitive prices. Inspired by him, I will apply the Decision-Making Theory to a real-life bidding situation.

On June 19, 2001 a translation job was advertised on proz.com website. For many projects the agencies keep the biding private but this was a public bid. Some translators bid on the job on a word-count basis; others submit the total price they ask for the project. I converted all the bids into a total dollar figure in order to be able to compare them. As the agency indicated in the ad, this project concerned the transcription of a ten-hour audio record in Turkish and its translation into English. Looking at the bids submitted, to which bidder should the agency grant the project?

Before we try to solve the problem, we need to make a few conversions in order to make the data more understandable. When we assume that we speak at a rate of approximately 100 words a minute, the total word count of the project in the source language would be about 60,000 words. Full-time professional translators can produce 2000-2500 words of quality translation a day (if they do not work overtime). So at least 25-30 days will be needed for translation and a few days for transcription of the source text. The deadline given by the agency is July 21, which is 32 days from June 19. It is doable.

I converted the word-count-based bids into approximate total dollar figures based on a 60,000 word count assumption. The resulting table of bids is presented below:

Let us start with the elimination process before we try to run the AB/QM program for this data in order to see which is the best bid. As can be seen above, some translators did not specify either the time they would need to complete the work, or the amount money they are asking for the project. When an agency receives 27 bids in one day—and some of them are outrageously low—it is wishful thinking on the translator's part that the agency will make an international or long-distance call just to ask when he/she could finish the project if he/she gets the job or how much she/he would ask for it. Obviously bidders # 2, 3, 9, 13, 14, 17, 18 submitted insufficient information and they would be eliminated from the evaluation process immediately. It is also appropriate to do so because AB/QM would not allow us run the program otherwise. Bid # 24 is also eliminated because #24 and #23 are the same bids submitted by the same person twice. So now we only have 18 bidders:

As we can see from the table, the bids range from $500 to $5400 and 2 days to 32 days. When we run AB/QM by using the decision-making theory and its cost minimization alternative, the software tells us that the best choice for the agency to choose is bidder #10. Bidder #10 offers the lowest price and fastest output. This choice is also obvious when we look at the table. Unfortunately this might not be the most logical decision to make on the agency's part. We need to make sure it is the best choice even under the worst-case scenario. Let's consider this possibility.

My three years' experience in freelance translation taught me that under normal circumstances a professional full-time (8 hours a day) translator who is fluent in both languages and familiar with both cultures can produce approximately 2000 words of quality translation in day. This would mean that the translator would need approximately 240 hours to translate 60,000 words. Bidder #10 tells us that he/she could transcribe and translate the entire document in seven days, which is simply impossible unless multiple translators are employed. He/she might be thinking about employing 3 or 4 translators in order to meet the deadline, but this solution is unfeasible for the individual translator. This means that if this translator is picked for the project, the quality of the translation will suffer considerably because of the money and time constraints he/she will have to deal with. If this is a risk that the agency is willing to take, they should pick bidder #10. The question we need to answer at this point is: Is this risk worth to take?

Let us try to find out what the agency could lose if bidder #10 is selected for the project. In order to do so, we need to assume the worst-case scenario and retest the above result.

After the documents are translated they should go through editing and proofreading processes by the second and third language professionals. These two processes also add extra expenses and time to the process. A professional editor could edit (compare the translation to the original document) approximately 2000 words of a good-quality translation in an hour and approximately 4000 words could be proofread (for punctuation and small grammatical errors) in an hour by a skilled proofreader. Most proofreaders and editors work on an hourly fee basis in translation business. A good-quality translation would minimize the agency's expenses and time lost in the editing and proofreading processes while a poor-quality translation would extend the time both the editor and the proofreader will spend on the project and add to the cost.

Again, from my own experience, I also know that the unit market price for this kind of good quality translation in the English-Turkish language pair is $0.15 a word. An hourly wage for editing is $30.00/hr and $20.00/hr for proofreading. So, under normal circumstances, if the agency wants the best and fastest translation it should spend:

60,000 words x 0.15=$9000 (translation)

(60,000 ÷ 2000) x 30,00=900 (editing at $30.00 per hour)

(60,000 ÷ 4000) x 20,00=300 (proofreading at $20.00 per hour)
______________________________________

TOTAL: $10,200

Anything below this figure would be pure profit for the agency. From these figures we could assume that in order for the agency to maximize its profit, the logical thing to do is to minimize the translation even at the expense of increased editing and proofreading time—which will increase both editing and proofreading cost for the agency. Let us assume that the poor translation would triple the time spent in editing and proofreading, and a fair translation would double the time spent in both editing and proofreading. Based on this assumption and the regular editing and proofreading prices stated above, let's calculate the cost of a good, bad and fair translation to the agency for each bid and run the AB/QM:

AB/QM Output:

Program: Decision Theory/Decision Making Under Risk

Problem Title: The best alternative/second test of the data

********Program output********

Expected Cost Table

<= indicate (s) the best alternative (s)

*******End of the output********

As the AB/QM program output above indicates, the best bid for the agency is still the bid #10. Even under the worst-case scenario (if the translation quality is bad and it triples the editing and proofreading time and cost) it leaves the largest profit to the agency and it still is the quickest process.

Although it may seem uncomplicated, the solution of the problem presented above required many mathematical calculations, and assumptions. Assumptions are made from a subjective point of view and are based on my personal experiences. At the end the AB/QM software indicated that the lowest bid with a shortest deadline was the best choice for the agency. Although the result may seem relatively clear and simple, I still feel a little uncomfortable making this decision. Although I tried to eliminate the effects of the poor-quality translation and what it could cost the agency, I still feel it does not lend itself to mathematical treatment. The correlation between the time spent in translating and the quality of the translation is still unknown. It is highly possible that they might have a strong positive correlation. As the time spent on translation increases, the quality of the translation may also increase. If this is the case, the safest bet for the translation agency could be the lowest bid and longest completing time, bidder # 12 ($900 and 29 days). Unfortunately, it is hard to guess the intentions of the agency. They might be operating under a very tight deadline (which is almost always the case). The short deadline could be the primary reason for opening this job to bids, in which case they can neither tolerate the lengthy editing and proofreading processes for a bad translation nor can they wait for the cheap but good translator to complete the job in 29 days. It would be interesting to know which bid the agency chose based on which criteria and what quality of translation they end up receiving. Without knowing the specific criteria the agency is using under certain circumstances, any biding game in freelance translation business looks like guesswork to me, even if the bidder takes a logical and mathematical approach to it. There might be some statistical criteria exploring the relationship between production time and quality in certain industries. There may even be some research about how aggressive competition obscures this statistical relationship between quality and production time. But I doubt such statistical data or research exists in the translation business. For that reason, I suggest for the future research to focus on collecting data from multiple sources and exploring the relationship between these two variables in order to be able develop more comprehensive methods for decision-making processes.